We propose a 4-year continuation of the Collaborative Longitudinal Personality Disorders Study (CLPS), a naturalistic, prospective, multicenter study of 733 treatment-seeking subjects. We will obtain a minimum of 10 years of follow-up on subjects (6 years on recently recruited minority subjects) diagnosed with four DSM-IV personality disorders (PDs): schizotypal, borderline, avoidant, and obsessive-compulsive. In addition, we will test a "hybrid" model of PDs that consists of more stable personality traits linked to intermittently expressed, symptomatic behaviors. Our findings to date document that PDs show consistency as syndromes over time, but rates of improvement that are inconsistent with DSM-IV definitions. Our central goal for the renewal period is to develop further an empirical base for redefining PDs in terms of their more vs. less stable elements. This requires a focused follow-along assessment to continue to describe (Aim 1) and to predict (Aim 2) the course(s) of PDs in contrast to major depressive disorder (MDD) and to track the temporal course and mediators of change of both stable elements (traits) and unstable elements (symptomatic behaviors) of PD psychopathology (Aim 3). Subjects will be assessed annually with instruments focused on the essential components of PD psychopathology and their course mediators, including trait and symptom status and severity, course of Axis I disorders, psychosocial functioning, treatment utilization, and stressful life events. Enrollment of an additional 65 African-American and Hispanic subjects with PDs during the past 4 years now enables us to examine these in expanded ethnic minority populations. CLPS is unique in its large number of personality-disordered subjects, comprehensiveness of their assessment, and length of prospective follow-up. This proposal will result in meaningful new knowledge about personality psychopathology with direct implications for classification, research on underlying phenotypes and mechanisms of change, and the development of new targets for treatment.